1. Field of the invention
The present invention relates to a system and method to give a true indication of respondent satisfaction to an electronic questionnaire survey from a plurality of respondents. More particularly, but not exclusively, the invention relates to a system and method for companies to accurately measure employee satisfaction and customer satisfaction efficiently and reliably.
2. Description of Related Art
Before the advent of computers, the only way to conduct a questionnaire survey was by using a traditional pen and paper method. This tried and trusted approach asked respondents a number of questions to which the answers could be a simple yes/no, a multiple choice type of response, a multiple selection type of response or a value on a scale (for example ranging from ‘I fully agree’ through ‘I don't have an opinion’ to ‘I don't agree’). By using various techniques such as repeat questions, it was also possible to qualify the respondent's answers to some extent.
For this type of questionnaire survey, an assessment of the results is typically based upon a statistical analysis of the answers given. Whereas this is a useful tool and much can be read into such a statistical analysis, the actual interpretations of the results can vary, depending on the assessor. Such a system is also open to abuse: the author is aware of situations where such an analysis has been interpreted in order to meet a customer's expectations - something which defeats the whole object (and cost) of the exercise.
The development, deployment, collation and assessment of such pen and paper questionnaire surveys were, and still are, typically highly labour intensive, even for small sample sizes, and as a result expensive.
With the introduction of computers and also Wide Area Networks (WANs) it became possible to harness the computer's power to alleviate some of the manual work in the whole questionnaire survey process by using, as an example, email and/or the internet to distribute and collate questionnaire surveys. Despite being able to simplify certain parts of the questionnaire survey process the process still remained highly labour intensive at the survey creation and survey analysis stages.
As computers became more powerful they also became easier to use (due to a plethora of diverse software being available to simplify the majority of daily tasks across all walks of life) impacting electronic versions of questionnaire surveys also. Not only were Internet based questionnaires becoming easier to manage and thus produce, but also so-called ‘kiosk’ survey systems were being set up (using dedicated standalone input systems in public places to attract the general public to take part in a survey).
Today there are many diverse applications for generating electronic questionnaire surveys available from the simple questionnaire surveys available free of charge on the Internet through to more complex software product offerings, where techniques such as ‘branching’ (the questions being asked being dependent on previous answers given) and ‘inconsistency testing’ (to determine whether questionnaire responses derive from a computer, or from a human not paying attention) are used to qualify the responses obtained.
Despite the large number of alternative choices available today for conducting an electronic questionnaire survey, all require the data to be analysed once the inputs have been collated. Due to the very nature of an electronic questionnaire survey this is almost always based on statistics, requiring interpretation by a human. Herein lies the fundamental problem:
Problem Analysis:
As mentioned previously, one of the main problems with depending on a statistical analysis is that it can be interpreted in numerous different ways, with each way having a different implication. As such, a statistical analysis only becomes useful when conducted more than once and where the trends between each analysis are measured, ensuring each time that the same analysis techniques are used. To get the most out of a questionnaire survey, therefore, requires either running the same questionnaire a number of times over a certain time period and monitoring the trends, or running the same questionnaire survey numerous times in parallel and comparing the results, ensuring each time that exactly the same technique is used to interpret the results.
However, there are several reasons why questionnaire surveys can not be run so frequently: firstly, as has been mentioned before, the costs involved can be prohibitive; secondly, finding people to survey and continue surveying becomes increasingly difficult with time—any ‘novelty effect’ wearing off very quickly—and using incentive programmes (a technique commonly used to increase the number of willing survey participants) also becomes expensive with time.
A second issue concerning questionnaire surveys specifically designed to measure respondent satisfaction (although not exclusively), is the time factor. Respondent satisfaction is very time dependent—one only has to monitor political opinion poll results to verify this. If a survey takes any length of time to complete, collate and analyse, the results thus obtained may no longer necessarily reflect the sentiments of those respondents at the time of the final result (although they were representative when the questionnaires were complete).
A third issue concerning questionnaire surveys specifically designed to measure respondent satisfaction (although not exclusively), is the fact that such questionnaires are answered emotionally rather than rationally and are therefore unlikely to be a true reflection of respondent satisfaction (emotions, being based on human feelings, are never stable, whereas satisfaction is an average measure of emotions over time). Also, asking a respondent to state their opinion (on which satisfaction is based) in a questionnaire is extremely difficult due to opinions being inherently analogue in nature, and trying to measure an opinion in an analogue fashion (from negative to positive in 10 steps, say) is also difficult to interpret (is the respondent who answers a question requesting an opinion on a subject in hand with −4, for example, really more dissatisfied than a respondent who answers the same question with −2?).
A final issue concerning questionnaire surveys specifically designed to measure respondent satisfaction (although not exclusively), is in summarising the results: If the same question is asked to two different respondents, where the response to be given is based upon a ten step scale (as discussed above) and one answers with +2 and the other with +8, summarising these results would give a value of +5. This may not, however, truly represent the combined feeling of both respondents. If, for example, the respondent with the +8 response was more convinced about their response than the respondent with the +2 response, then a true representation of the summary of these inputs would be expected to be +6 or +7. Such a result can only be obtained in a summary if the ‘level of conviction’ of the respondent is also taken into account in the results and therefore the end calculation—something which current survey techniques do not accomplish.